Chandra and HST observations of gamma-ray blazars

31 Pages·2007·0.33 MB·English

SLAC-PUB-12402
astro-ph/0703359
March 2007
Chandra and HST observations of gamma-ray blazars: comparing
jet emission at small and large scales
F. Tavecchio
INAF - Osservatorio Astronomico di Brera, via Bianchi 46, 23807 Merate (LC), Italy
L. Maraschi, A. Wolter
INAF - Osservatorio Astronomico di Brera, via Brera 28, 20121 Milano, Italy
C. C. Cheung1
Kavli Institute for Particle Astrophysics and Cosmology, Stanford University, Stanford, CA
94305, USA
R.M. Sambruna
NASA Goddard Space Flight Center, Code 661, Greenbelt, MD 20771, USA
C.M. Urry
Yale Center for Astronomy and Astrophysics, Yale University, 260 Whitney Avenue, New
Haven, CT 06520, USA
ABSTRACT
We present new Chandra and HST data for four gamma-ray blazars selected
on the basis of radio morphology with the aim of revealing X-ray and optical
emission from their jets at large scales. All the sources have been detected.
Spectral Energy Distributions of the large scale jets are obtained as well as new
X-ray spectra for the blazar cores. Modeling for each object the core (sub-pc
∼>
scale) and large-scale ( 100 kpc) jet SEDs, we derive the properties of the
same jet at the two scales. The comparison of speeds and powers at different
scales supports a simple scenario for the dynamics and propagationof high power
relativistic jets.
Subject headings: Galaxies: active — galaxies: jets — (galaxies:) quasars: indi-
vidual (0208-512, 0954+556, 1229-021, 2251+158) — X-rays: galaxies
1
Jansky PostdoctoralFellow of the National Radio Astronomy Observatory.
Submitted to Astrophys.J.
Work supported in part by US Department of Energy contract DE-AC02-76SF00515
– 2 –
1. Introduction
In a unified scenario, relativistic jets originating from the accreting black hole and
propagating outwards to kiloparsec scales and beyond must be present in all radio-loud
active galactic nuclei (AGN). How these jets form and evolve is not yet known; even intrinsic
jet power and composition are only poorly known. Thus jets remain one of the key puzzles
of AGN astrophysics (see e.g. Blandford 2001, De Young 2002).
The discovery of X-ray emission from kiloparsec-scale jets – a major success of Chan-
dra (see Harris & Krawczynski 2006 for a review) – has provided us with a new tool to
probe jet physics. Dozens of new X-ray jets have now been found (see the updated list at
http://hea-www.harvard.edu/XJET/ and references therein), and associated Hubble Space
Telescope (HST) observations have led to a near doubling of the number of known optical
jets (a list can be found at http://home.fnal.gov/~jester/optjets/).
The origin of this large-scale emission is probably complex. Extended X-ray emission
from low power (FRI) jets is likely synchrotron radiation from very high energy electrons
connecting smoothly ina single emission component tothe electrons responsible for theradio
emission(e.g.,Worralletal. 2001,Kataoka&Stawarz2005). Ontheotherhand,thespectral
energy distributions (SEDs) of the multifrequency emission from high power jets in quasars
generally require two spectral components, as was the case for the first discovered extended
X-ray jet (PKS 0637-752; Schwartz et al. 2000). For instance, in the X-ray survey of selected
radio jets by Sambruna et al. (2004) which discovered 34 emission regions in 11/17 jets, only
3 are consistent with a single (power law, or convex) synchrotron component while for the
other emission regions the SED has a concave shape (see also Schwartz et al 2006). The
“rising” X-ray spectral component has been successfully interpreted in a number of cases as
inverse Compton (IC) scattering of boosted CMB photons by low energy electrons in the jet,
implying highlyrelativistic bulkmotionuptovery largescales (Tavecchio et al. 2000,Celotti
et al. 2001, Siemiginowska et al. 2002, Sambruna et al. 2004,2006a, Cheung 2004,Tavecchio
et al. 2004, Schwartz et al. 2006). At this point, other models cannot be discarded, and
it is also the case that some problems in applying the IC/CMB model to specific sources
have been pointed out (see Stawarz et al. 2004, Atoyan & Dermer 2004, Kataoka & Stawarz
2005, Uchiyama et al. 2006, Jester et al. 2006, Harris & Krawczynski 2006 for criticisms and
alternatives). At the same time, the one-zone IC/CMB model, if correct, has the advantage
of involving few physical parameters which can be constrained by the radio, optical and X-
ray emission of kiloparsec-scale jets, yielding interesting consequences on the physical state
of the plasma in the jet and on their dynamics at large scales (e.g. Sambruna et al. 2004,
2006a, Marshall et al. 2005, Schwartz et al. 2006).
At the opposite extreme of spatial scales, the innermost region of jets is fruitfully in-
– 3 –
vestigated through the study of blazars, whose emission is interpreted as the relativistically
amplified non-thermal continuum produced close to the central engine (d ∼ 0.1 pc) by a
jet closely aligned to the line of sight (Urry & Padovani 1995). Spectral modeling of the
full infrared through γ-ray SED (with synchrotron + IC components) yields the electron
density and energy distribution, magnetic field intensity and bulk Lorentz factor of the flow
(with some confidence, as shown by the agreement between results of different groups; e.g.,
Ghisellini et al. 1998, Tavecchio et al. 2002, Kubo et al. 1998, Sikora & Madejski 2001),
allowing us to infer basic global quantities characterizing the jets quite close to their origin,
in particular the kinetic power and the matter content (e.g., Maraschi & Tavecchio 2003).
Coupling information derived at subparsec- and kiloparsec-scale for the same jet could
have great potential to construct a global understanding of powerful extragalactic jets. This
approach was first applied to two well known blazars (1510-089 and 1641+399) serendip-
itously belonging to the sample surveyed with Chandra by Sambruna et al. (2004). The
sample had been selected on the basis of radio morphology and brightness in order to search
for extended X-ray emission. For both blazars X-ray emission outside the nucleus was de-
tected. The results (Tavecchio et al. 2004, hereafter Paper I) were consistent with a constant
bulk Lorentz factor and constant power along the jet and a simple scaling of the electron
density and magnetic field intensity, suggesting free expansion of the jet between the subpc
∼
(blazar) scale and the ( 100 kpc) scale of the resolved X-ray knots. Therefore, these results
imply that powerful jets are only weakly affected by the environment up to these scales.
Note, however, that these findings do not exclude that the terminal portions of the jets can
interact more strongly with the environment, dissipating part of their power and deceler-
ating. Indeed, there is evidence (e.g. Georganopoulos & Kazanas 2004; Sambruna et al.
2006a) suggesting that, in some cases, the terminal regions of the jet can be affected by de-
celeration (possibly due to the cumulative effects of entrainment of external gas; Tavecchio
et al. 2006), accompanied by an increase of magnetic field intensity and particle density.
Unfortunately, only very few jets can be studied on both scales. Indeed the best-studied
inner jets do not tend to have well studied large-scale jets, precisely because the former
are the most closely aligned with the line of sight, which means projected jet lengths are
small and the large scale jets must be seen in contrast to the bright, beamed cores. This
is why the number of blazars well observed on pc and kpc scales is very small. It is also
worth noting that in the case of sources displaying large scale jets but with a weak blazar
core, the determination of the parameters characterizing the small scale jet is less robust,
due to the presence of a mix of jet and disk emission in the core, especially in the crucial
X-ray band (e.g. Sambruna et al. 2006b). With the aim of increasing the number of
blazars with multifrequency observations of the jet at large-scales, we proposed Chandra
and HST observations of four γ−ray blazars (0208-512, 0954+556, 1229-021 and 2251+158;
– 4 –
see Tab.1), showing a radio jet suitable for X-ray imaging. The detection of the blazar in
γ-rays indicates the presence of a strong jet component in the core, generally leading to a
more reliable estimate of the small scale jet parameters due to the extended sampling of the
SED.
In this paper we report the analysis of Chandra and HST1 data (Sect. 2), the modeling
of the SEDs of the blazar cores and knots in the large scale jet (Sect. 3), the determination
of speeds and powers (Sect. 4) and the discussion of the results (Sect. 5). Throughout this
work we use the following cosmological parameters: H = 70 km s−1 Mpc−1, Ω = 0.7,
0 Λ
Ω = 0.3.
M
2. Observations and data analysis
Basic characteristics of the blazars analyzed in this work are reported in Table 1. In the
following we report the details of the procedure used to analyze the data. For the quasar
0208-512 we did not obtain the requested Chandra pointing, since it was assigned to the
jet survey of Marshall et al (2005). X-ray and radio data for the core for this source are
directly taken from Marshall et al. (2005) and Schwartz et al. (2006), while we report only
the analysis of our HST pointings.
2.1. Chandra
Thesourceswereobserved withChandra(Weisskopf etal. 2000)withACIS-Sinimaging
mode. The journal of observations is reported in Tab.2. The data were collected with the
back-illuminated ACIS-S S3 chip in 1/8 sub-array mode to avoid/minimize the pileup of the
central AGN source, with a frame time of 0.4 s. Telemetry was in faint mode.
The data were reduced with the standard Chandra pipeline with the CIAO software
(version 3.3) and the most recent available calibration products (CALDB 2.26), as described
in http://cxc.harvard.edu/. The corrections applied are those appropriate for the ACIS-
S instrument. No high background periods due to particle-induced flares were present in the
datasets.
Events were selected in the 0.3-10 keV interval for both imaging and spectral analyses.
1BasedonobservationswiththeNASA/ESAHubbleSpaceTelescopeobtainedattheSpaceTelescopeSci-
enceInstitute,whichisoperatedbytheAssociationofUniversitiesforResearchinAstronomy,Incorporated,
under NASA contract NAS5-26555.
– 5 –
The radio and X-ray images were registered assuming that the cores are coincident. Applied
′′ ′′ ′′
shifts are 0.42 , 0.22 , 0.32 for 0954+556, 1229-021 and 2251+158, respectively, below
′′
Chandra aspect uncertainties. The core spectra were extracted in a 1.5 region centered on
the centroid of the emission. The jet spectra were extracted in circles centered on bright
radiofeatures(seebelow). Thebackgroundwasevaluatedinnearbyregionsdevoidofsources.
Response matrices are created in the usual fashion; spectra are binned so that each resulting
bin has a S/N > 3.
2.1.1. Cores
Due to the short frame time the pile-up is negligible in both 0954+556 and 1229-021.
Fitting the spectra with a power-law model and free N gives a value consistent with the
H
Galactic column density (from Dickey & Lockman 1990). We therefore fit the spectra with
fixed galactic N and a simple power law. Results are given in Table 2. 2251+158 is at
H
least a factor of 10 brighter than the other two targets, and the effects of pile-up cannot
be ignored (pile-up fraction=29%). We use the pileup model in XSPEC (Davis 2001) to
account for the distortion in the spectrum. In this case we find an N slightly in excess of
H
the Galactic value.
These sources have been observed several times in the past in X-rays (Fossati et al.
1998, Wilkes et al. 1994, Siebert et al. 1998, Prieto 1996, Tavecchio et al. 2002, Marshall et
al. 2005) with fluxes and spectral parameters consistent with our results.
2.1.2. Jets
Jet knots are quite fainter than the cores and the poor statistics do not allow to measure
the X-ray spectral slope. Fluxes listed in Table 3 are derived assuming a power-law spectrum
of slope Γ = 1.7 (typical for jet knots, e.g. Sambruna et al. 2006a) and Galactic N .
H
2.2. HST
We observed three of the four blazars with the Advanced Camera for Surveys (ACS)
aboard the Hubble Space Telescope in two bands. The remaining source, PKS 1229–021,
has existing multi-filter WFPC2 data available from the archive from which Le Brun et al.
(1997) already detected optical emission in the jet. For our ACS observations, we observed
each target with the F475W and F814W filters (SDSS g and Broad I, respectively, with
– 6 –
effective frequencies of 6.32×1014 Hz and 3.72× 1014 Hz) for one orbit per filter. The Le
Brun et al. (1997) data were taken in the F450W and F702W filters, close to standard B
and R bands, respectively, with effective frequencies of 6.58×1014 Hz and 4.33×1014 Hz.
In 2251+158, Cheung et al. (2005) discovered optical emission from the jet and hot
spot from an archival WFPC2 image. The new ACS images (characterized by a better
resolution) confirm the detections of knots B and C in the jet (nomenclature from Cheung et
al. 2005)andresolve thehotspotinto2pieces –thisperhapsmeansthatthereisapreviously
unresolved portion of the jet in the “hot spot” as reported in Cheung et al. (2005). Knots A
andBarenoteasilydistinguishablefromthewingsofthebadlysaturatedopticalnucleus and
wecouldnotobtainanyusefulphotometryofthemfromtheACSimages. Ourmeasurements
of the optical jet fluxes in PKS 1229–021 are consistent with those published in Le Brun et
al. (1997).
Upper limits of 0.03 µJy (3 σ, in both the F814W and F475W images) were measured
for radio jet knots undetected in the ACS images (0208–512 and 0954+556). In the archival
WFPC2 images of PKS 1229–021, the corresponding point sources limits are 0.2 and 0.3 µJy
(3 σ) in the F702W and F450W images, respectively.
2.3. Radio
For 3 of the 4 targets, we obtained and analyzed radio data from the NRAO2 Very
Large Array (VLA) archive using standard procedures in AIPS (Bridle & Greisen 1994) and
DIFMAP (Shepherd, Pearson & Taylor, 1994). We use published ATCA measurements of
PKS 0208–512 from Schwartz et al. (2006).
A single 4.9 GHz image of 0954+556 was analyzed (3.5 min. from program AH170) to
set the normalization for its extended radio emission; Reid et al. (1995) measured a radio
spectral index for the western jet/lobe complex of α=0.9 using data from 0.4–5 GHz and
we adopt this value. For PKS 1229–021, matched resolution 4.8 and 15 GHz data originally
published by Kronberg et al. (1992) were reanalyzed. We use published measurements and
dataforthejetof2251+158fromCheung etal. (2005). Forthelattertwo targets, wecreated
spectral index maps which show values of α ∼0.7 (PKS 1229-021) and α ∼0.8 (2251+158)
for the jet.
2The National Radio Astronomy Observatory is operated by Associated Universities, Inc. under a coop-
erative agreement with the National Science Foundation.
– 7 –
2.4. Results: imaging and photometry
In Figs. 1 (a-c) we show the smoothed X-ray images together with radio contour over-
lays. TheChandraimages(plottedinlogarithmicscale)havebeensmoothedwiththeftools
tool fadapt using a threshold of 10 counts for 0954+556 and 1229-021 and 20 counts for
2251+158. Radio contours are plotted logarithmically in steps of a factor 1.5.
X-rayemissionassociatedtobrightradioknotsisclearlydetectedinallthethreesources.
In0954+556,duetothelimitedlengthofthejet, onlytheemission associatedtotheterminal
region can be clearly evaluated. In the jet of 1229-021, X-ray emission tracking the radio is
clearly visible. The brightness of the X-ray emission decreases and is only barely detected
at the terminus. One knot and the hotspot are detected in 2251+158. In Figs we report
the circular region used to extract the fluxes. Table 3 summarizes the values of the fluxes
measured in the extraction regions centered on bright X-ray features (for which we use an
alphabetical nomenclature, starting with the region closest to the core)
3. Modeling the Spectral Energy Distributions
In Fig. 2 (a-d) we report the SEDs, for the cores (upper panels) and emission regions
in the large scale jet (lower panels), constructed using data from the historical records and
the Chandra and HST data presented in this work. Historical data used to construct the
blazar SEDs are taken from the references reported in Tavecchio et al. (2002) for 2251+158
and 0208-512, while those for 0954+556 and 1229-021 are reported in the figure caption. It
is worth noting that the blazar nature of 0954+556 and its identification with the EGRET
source 3EG J0952+5501 has been recently questioned on the basis of new VLBA images
revealing a Medium Symmetric Object morphology (Rossetti et al. 2005; see also Marscher
et al. 2002). This is contrary to the typical morphology of blazars, characterized by the
presence of strong, compact features. It is therefore more likely that the counterpart to
3EG J0952+5501 is J0957+5522 (Sowards-Emmerd et al. 2003), not 0954+556 as originally
suggested by Mattox et al. (2001).
All these blazars are well studied, with several observations in the X-ray band. For clar-
ity we only report the new Chandra data. We stress that all these data are not simultaneous
and that the sources can undergo large variations. In this respect, the best example of a
variable source among those considered here is 2251+158, which in the spring 2005 displayed
a strong outburst, with a large (at least a factor of 3) increase of the optical and the hard
X-ray luminosities with respect to quiescent levels (Pian et al. 2006)
– 8 –
3.1. The Blazar region
We modelled the SEDs of the inner jet with the emission model fully described in
Maraschi&Tavecchio(2003),consideringsynchrotronandIC(bothSynchrotronSelf-Compton
and External Compton) radiation.
To reproduce the observed shape of the two humps in the SED we assume that the
electron energy distribution is described (between γ and γ ) by a smoothed broken
min max
power law with indices n and n below and above the break located at γ . This purely
1 2 b
phenomenological description accounts for the observed shape of the synchrotron and IC
components. In the widely assumed diffusive shock acceleration model (e.g. Kirk et al.
1998) or in cases of severe cooling one would expect n = 2. However, there are several
1
objects for which this limit seems to be violated, with values as extreme as n = 1.4−1.5
1
(e.g., Piconcelli & Guainazzi 2005, Yuan et al. 2005). It is conceivable that, at least in these
cases, the electron distribution derives from a (continuously operating) different acceleration
mechanism (for possibilities, see e.g, Sikora et al. 2002; Katarzyn´ski et al. 2006).
Relativisticelectronsandtangledmagneticfieldwithintensity B fillthesource, assumed
to be a sphere with radius R. Relativistic effects are described by the relativistic Doppler
factor δ, given by δ = [Γ(1−βcosθ)]−1, where β = v/c, Γ is the bulk Lorentz factor and θ
is the viewing angle with respect to the blob velocity.
In the SED we also include a black body with luminosity L and temperature T = 104
disk
K, intended to provide a crude representation of the blue bump originating in the accretion
disk. A fraction L = τL of this radiation is thought to be reprocessed by the Broad
BLR disk
Line Region clouds, located at a distance R from the central BH, providing the source
BLR
of the external photons for the EC process.
Estimates of L for 0954+556, 1229-021 and 2251+158 are given by Cao & Jiang
BLR
(1999). For 0208-512 we estimate L using the methods of Celotti et al. (1997) and
BLR
the flux of the MgII line provided by Scarpa & Falomo (1997). Given L , we fix L
BLR disk
assuming the parameter τ fixed to 0.1, a value consistent with the ratio between L and
BLR
L inferred for the (few) radio-loud sources for which a measure of both quantities is
disk
simultaneously available (e.g. Sambruna et al. 2006b).
In Fig. 2 (top panels) we report the total emission (solid line) and the separate con-
tributions from the different spectral components (synchrotron: dotted; SSC: short dashed;
EC: long dashed; disk: dot-line). In all cases the measured slope of the X-ray continuum
appears to be hard but not as extreme as in other powerful blazars (where α < 0.5, e.g.
X
Tavecchio et al. 2002), suggesting an important contribution of the (soft) SSC component
which, typically, peaks close to the X-ray band.
– 9 –
The parameters used to calculate the model (reported in Tab. 4) are similar to those
usually found for this type of sources (e.g. Maraschi & Tavecchio 2003). In all cases we fix
the value of the minimum Lorentz factor of the emitting electrons at γ = 1, as derived in
min
the case of other powerful blazars (e.g. Maraschi & Tavecchio 2003). However, this was not
possible in the case of 0954+556, because of the relatively steep X-ray continuum (α ∼ 1),
X
which favors a dominant contribution of the SSC emission and a minor contribution from
the (flatter) EC component. In this case we use γ = 8. However, we recall that the
min
association with the EGRET emission is questionable, and therefore the results for the
latter source should be considered with caution. For all the sources we used a value of R
BLR
consistent with the the correlation between R and L found by Kaspi et al. (2005)
BLR BLR
and Bentz et al. (2006).
Although the sampling of the SEDs is good only in the case of 2251+158, the data
constrain the model parameters sufficiently for the present purpose. In fact the lack of data
for some of the objects in the 1011−1014 Hz region leaves rather unconstrained a portion of
the synchrotron continuum which does not influence the derivation of the most important
parameters. In particular, the power critically depends on the number of particles carried by
the flow (see Sect. 4), which, due to the steep electron distribution, is constrained by the well
defined X-ray emission, dominated by EC radiation from low-energy electrons. Note also
that the intensity of the γ-ray emission determines the radiative output (i.e. the radiative
efficency) of the jet but has a limited influence on the derivation of the kinetic power, since
γ-ray emitting electrons do not contribute significantly to the total number of particles.
3.2. The Large-scale jet emission
We model the emission using the IC/CMB model, already used in Paper I. The emitting
region is modeled as a sphere with radius R, filled by relativistic electrons and tangled
magnetic field with intensity B, in motion with Lorentz factor Γ. Electrons follow a power-
law law in energy with index n, N(γ) = Kγ−n, between the extremes γ and γ and emit
min max
through synchrotron and Inverse Compton emission. For the latter it is assumed that the
dominant soft radiation field is the CMB. To uniquely determine the parameters we assume
equipartition between the relativistic electrons and the magnetic field (see Sambruna et al.
2006b for a discussion).
The radius assumed in the models corresponds to the angular size of the circular re-
gions used to extract the fluxes. As suggested by the radio images, with this choice we are
probably overestimating the actual emitting volume. However, it is easy to show that the
derived parameters depends rather weakly on the assumed volume (see e.g. the Appendix of
– 10 –
Tavecchio et al. 2006). For instance, with a the volume equal to 1/10 of the value assumed
here, both the magnetic field and the Doppler factor would be larger by a factor ∼ 1.3.
Similarly, the impact of the assumed size on the derive power is rather minor. In the case of
0208-512, for which Schwartz et al. (2006) use a rectangular extraction region (and therefore
model the emission region as a cylinder), we derive an effective radius, such that the volume
of the sphere is equal that of the cylindrical region. Apart for the case of 0208-512 we fix the
slope of the electron energy distribution using the radio spectral index derived above (Sect.
2.3). The 0208-512 radio jet spectral index of 0.8 is assumed, as the actual measurements
over the short frequency baseline (4.8-8.6GHz) gave unreliable values (Schwartz et al. 2006).
The value of the minimum Lorentz factor of the electrons has been chosen so that the break
in the IC continuum is located between the optical and the soft X-ray band. Of course the
choice of this value is not unique, since values of γ in the range 5–30 are usually allowed
min
by the data. A deeper discussion of this point is reported in the next paragraph.
In Fig. 2 (lower panels) we show the synchrotron-IC/CMB spectral models derived
assuming the input parameters reported in Tab.5.
For 0954+556 the X-ray flux could only be extracted for the region corresponding to the
terminalportionofthejet. Asdiscussed inTavecchio etal. (2005),alsointhiscaseabeamed
IC/CMB model appears necessary, however with a Doppler factor lower than typically found
for knots in the jet.
For 1229-021 we can model the emission at three different locations along the jet. In-
terestingly, we can reproduce the SEDs of the three regions decreasing only the value of the
Doppler factor (from δ ≃ 9 to δ ≃ 5), and maintaining all the other quantities (almost)
constant. As well, the clear bending of the jet between regions A and B suggests a change
in the (projected) direction of the jet speed. It is tempting to associate this change with an
increase of the jet angle, which could simultaneously explain the decrease of the observed
beaming (while a “true” deceleration would imply also the increase of both the magnetic
field intensity and the particle density).
4. Speed and power
The parameters derived by reproducing the SEDs can be used to infer the bulk Lorentz
factor and the total power characterizing the jet. In the case of large scale jets with more
than one emission region, we calculate the power and the speed at the region closest to
the core. With this choice we avoid the portion of the jet near the terminal region which
could be affected by deceleration. For the same reason we do not include 0954+556 in this